Preeclampsia and preterm delivery are important causes of maternal and infant morbidity and mortality. Recent studies suggest that elevated air pollutant exposures may lead to preterm delivery, even at air pollutant levels typical of those in many US cities. No prior studies have directly examined air pollutant exposure in relation to preeclampsia; although much evidence suggests that the relation is biologically plausible. This study is based on a source population of women participating in a large cohort study. Participants enroll during early pregnancy and are followed until delivery. We will design models that use local traffic, weather, and population characteristics to predict monthly ambient concentrations of fine particulate matter (PM2.5) and carbon monoxide (CO). These models will be used to estimate study participants' PM2.5 and CO exposures before and during pregnancy. We will test whether these air pollutant exposures are associated with subsequent risk of preeclampsia and preterm delivery using multivariable logistic regression procedures. Additionally, we will test biological markers of maternal lipid peroxidation (thiobarbituric acid reactive substances) and inflammation (high sensitivity C-reactive protein) in maternal blood samples drawn during early pregnancy. We will use multivariable regression to examine lipid peroxidation and inflammation as biological mechanisms in air pollutant/disease relationships. We will also examine carboxyhemoglobin measured in early-pregnancy maternal blood samples as a marker of CO exposure. We will test the associations between carboxyhemoglobin and the pregnancy outcomes. Study results could contribute to knowledge regarding the causes and mechanisms involved in preeclampsia and preterm delivery. Such knowledge may have significance in developing preventative interventions for these two common adverse outcomes. The results could also contribute to the body of heath-related science that influences air quality standards and policies. [unreadable] [unreadable] [unreadable]